> We've had previous problems with ETFS where upon starting/mounting we got 
> DATAERRs and as a result fid truncations:
> 
> fs-etfs-mpc8313erdb512-nand512: readtrans DATAERR on cluster 474191
> fs-etfs-mpc8313erdb512-nand512: Truncating fid 0x18 from 16384 to 0 bytes
> 
> It seems we've been able to fix this, but now we've got a different type of 
> error: truncations without any previous DATAERRs! From the logfile:
> 
> fs-etfs-mpc8313erdb: readtrans BADBLK on cluster 507008
> fs-etfs-mpc8313erdb: Truncating fid 0x9 from 12288 to 2176 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0xb from 147456 to 2244 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0xc from 2048 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0xe from 2048 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x10 from 2048 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x14 from 2048 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x16 from 2048 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x17 from 57344 to 52232 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x18 from 47104 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x1a from 163840 to 0 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x1b from 260096 to 50304 bytes
> fs-etfs-mpc8313erdb: Truncating fid 0x1d from 2048 to 0 bytes
> 

When files are shortened, the new length is recorded in the .filetable file.  The old transactions that make up the 
portion of the file beyond the new EOF recorded in .filetable still exist (until a reclaim/wear-level cleans them up).  
During that window, if the ETFS driver is re-started, it will discover this discrepency, and issue a warning.  The file 
length recorded in the .filetable is always treated as the correct value, so the file itself is not corrupted, simply 
ETFS has discovered an old remenant of the file's previous contents.

ETFS has been changed to only display these messages when increased verbosity is requested.

> What can be the reason for those errors? ETFS exited after this.
> 
> Greetings,
>  Marc